Pathophysiological hallmarks of glioblastoma (GM), an incurable cancer with contemporary therapy, include microvascular proliferation and VEGF-mediated angiogenesis. We and others have demonstrated that interruption of VEGF-mediated signaling results in "normalization" of tumor vasculature, transforming disorganized, dilated, leaky vessels into a more organized blood vessel network with reduced diameter and permeability. Thus, normalization may enhance delivery of chemotherapeutics and render the tumor more susceptible to cytotoxic therapies via improved oxygen delivery and diminished tumor hypoxia. In this proposal we aim to extend observations from our prior study (1R21CA117079, Batchelor-PI) in which we demonstrated, via serial non-invasive MRI techniques, that AZD2171, an oral, pan-VEGF receptor tyrosine kinase inhibitor induces normalization of tumor vessels in recurrent GM patients. Normalization had rapid onset, was prolonged but reversible and had significant clinical and functional consequences. Increased blood levels of bFGF, SDF1 and circulating endothelial cells (CECs) correlated with tumor progression. This study was the first to identify the onset and duration of a vascular normalization "window" created by an anti- angiogenic agent in any human cancer and the first to demonstrate the potential utility of blood biomarkers as predictors of tumor progression on anti-angiogenic therapy. In this proposal we aim to apply the imaging and biomarker methods to patients with newly diagnosed GM. In Aims 1 and 2 of this proposal we will define, via application of dynamic contrast enhanced MRI;diffusion-weighted and perfusion-weighted MRI, the normalization window in newly diagnosed GM patients with the therapeutic goal of exploiting this window by combining AZD2171 with cytotoxic chemoradiation. We will determine the safety (6-12 patients) then efficacy (40 patients) of this combination in newly diagnosed GM patients. We hypothesize that this therapeutic strategy will improve survival;progression-free survival and radiographic response proportion in this patient population. Identification of the normalization window in GM patients may provide opportunities for the application of more intensive, time-dependent therapeutic strategies in the future. In Aim 3 we will continue to explore the potential utility of tumor and serial blood biomarkers, including bFGF, SDF1 and CECs, as predictors of tumor response and patient outcome in the newly diagnosed GM population.